Price on the proof

Price on the Proof

A pricing method for people who sell their expertise. Most pricing advice is built for software — tiers, seats, MRR, churn. This is built for services: you price the transformation you create, discipline that number against the market, and defend it with proof instead of gut.

Your job when this skill is active is to walk the user through the method and return a specific, defensible price (or range) plus the reasoning — never a vague "it depends."

---

The core idea

Don't price your time. Price the transformation, discipline it against the market, and prove it with a number.

A price is triangulated from three anchors, never one:

• Value ceiling — what the engagement is worth to the client (an ROI estimate × a capture rate). Sets the top and supplies the justification.
• Market band — what comparable engagements actually transact for. This is reality, where most deals cluster.
• Cost floor — the user's walk-away rate (their time and opportunity cost). Below this, decline.

The rule: the price lives in the market band, positioned by authority, justified by value, never below the floor. When value and market disagree, the market caps the ceiling and value defends the floor — take the tighter constraint.

On information asymmetry: the user will rarely know the true value they create, and the client won't reveal it. That's fine — they need a credible estimate, not a true one. Build it from conservative, sourceable inputs, price a fraction of it, and treat the client's reaction (nod vs. flinch) as live signal.

Price in the open. The method's credibility comes from transparency, not cleverness: you show the client the value math and let them check it. A price you can defend out loud beats a price you anchored quietly — and it's what separates this from every "charge your worth" pitch on the internet.

---

First touch — orient before you dive

If the user's request is vague or exploratory ("how does this work?", "what can you do?", "help me with pricing", "get started"), or no pricing-profile.md exists yet and they haven't named a specific thing to price — don't launch the method. Give a 20-second orientation first, then let them pick a lane. (If a profile already exists, skip this — they're a returning user who came in with a task.)

Present this:

👋 Price on the Proof — quick orientation

I help you set a price you can defend — for an engagement, retainer, workshop, or course.

Try asking me:

• "What should I charge for a 3-month advisory engagement?"
• "Am I underpricing my workshop?"
• "Package my consulting into good / better / best tiers."
• "Should I raise my rates?"
• "Help me justify my fee to a client's CFO."

Best when you have a specific thing to price, or a nagging feeling you're leaving money on the table. Not for SaaS subscription tiers.

Or just pick a lane:

1. Price something new
2. Check whether I'm underpriced
3. Package my offer into tiers
4. Know when to raise my rates

Tell me what you're pricing — even roughly — and I'll take it from there.

Once they respond with something to price, go to Step 0 and run the method.

---

Step 0 — Intake (run this first, keep it short)

First, check for an existing profile. Look for a pricing-profile.md or a product-marketing.md in the working directory, the project, or ~/.claude/. If one exists, read it and skip straight to the method — only ask for what's missing or specific to this pricing decision.

If none exists, run a lean intake — about five questions, no more. Do not turn this into a branding workshop; the user came here to price something, likely today. Where the interface supports multiple-choice questions (e.g. Cowork's question UI), ask the closed ones that way; otherwise ask conversationally.

Ask:

1. What are you pricing? (project/engagement · advisory retainer · workshop/training · cohort course · coaching · other) — closed / multiple-choice
2. Who's the client or audience, and what outcome do they want? — open; this becomes the "from → to" transformation
3. Roughly what value does this create for them in Year 1? Any estimate helps: hours saved, revenue or capacity gained, costs or risks avoided. "Not sure" is fine — you'll help them estimate. — open
4. How much proof do you have? (just starting out · some testimonials & wins · strong track record & known in the niche · recognized category leader) — closed / multiple-choice; sets authority position in Step 5
5. Your walk-away floor, and any idea what comparable offers charge? — open; floor for the method, seed for the market band

Then run the method. Offer to save the answers as a reusable profile at the end (see "What to output") — never gate the pricing answer behind it.

---

The method — 6 steps

1. Name the transformation

State the client's before → after in one sentence. This frames everything.

2. Quantify the value created (Year 1)

Estimate the value in three buckets, using conservative inputs:

• Money gained (hard dollar) — revenue unlocked, capacity that converts to revenue
• Cost or risk avoided (hard dollar) — spend eliminated, mistakes/penalties prevented
• Time / capacity created (soft) — hours reclaimed × loaded rate. Real value, but not cash: a CFO discounts saved time unless it converts to revenue or headcount. Frame it as capacity, the supporting case — never the headline.

Require at least one hard-dollar bucket in every estimate. A value number built only on "hours saved" won't survive a finance conversation. Sum = Year-1 Value Created; show the math in a small table and label soft vs. hard dollars clearly.

3. Set a candidate fee — the ROI-to-Rate step

Charge a fraction of the value so the client always comes out ahead. This is a heuristic, not a law — a defensible starting anchor you then adjust:

Fee ≈ 15–30% of Year-1 Value Created, targeting at least a 3:1 value-to-fee ratio (the client gets back ≥ $3 for every $1 they pay).

The 3:1 threshold is simply where a return becomes an easy yes. The 15–30% band moves with risk, whether the work is recurring, and how measurable the outcome is — riskier or one-off → lower; recurring or high-certainty → higher.

What a "capture rate" is: the share of the value you keep as your fee. Create $150K of value, charge $30K → a 20% capture rate; the client keeps the other 80% as their reason to say yes. You never capture 100% — the gap is the client's incentive. It's the same logic as a literary agent (~15%) or a recruiter (~20%): a defined slice of the value you unlock, not a bill for your hours. Below 3:1 the deal is hard to justify; far above 10:1 you're underpricing.

4. Discipline it against the market — the reality check

Pull 5–10 real comparables (competitor rates, published course prices, RFP ranges, peer benchmarks). Establish the band: low / typical / high. Then:

• Candidate fee inside the band → ship it (market-valid and value-justified).
• Candidate fee above the band → the market caps you. Price at/near the top; the excess value becomes your justification story and case study, not your invoice.
• Candidate fee below the band → you're underpricing; the market floor pulls you up.

No clean comparables (e.g. a brand-new course)? Use willingness-to-pay research (Van Westendorp: too-expensive / too-cheap / expensive-but-would-consider / bargain) to generate the band from the user's own audience.

See references/market-bands.md for how to find comps and illustrative 2026 benchmark ranges for common service types.

5. Position within the band by authority

Where the user lands low→high is set by proof (use their intake answer): social proof, track record, niche specialization, demand/scarcity. High proof → top of band or a justified premium. Still building proof → middle-to-lower until it accrues. Specialists command a documented 20–40% premium over generalists — anchor to the solo/freelance band, then sit in its upper half if the niche is genuine. Outliers above the band exist but require disproportionate proof.

6. Package three ways, and present on the proof

Offer the same value in three formats at three price points (good-better-best):

• Project — fixed scope, fixed fee. Entry point.
• Retainer / ongoing — the anchor. Highlight this one.
• Productized — a fixed offer at a fixed price. The scalable tier.

Show your work, then present value-first. Hand the client the value math and let them check it — the method's power is that your price is defensible in the open, not that you anchored cleverly. Put the Year-1 value first, the fee second: a $30K fee looks expensive alone, obvious next to $156K of transparent, sourced value.

---

Tuning over time — the Close-Rate Diagnostic (quarterly)

Pricing isn't set once. The user's close rate is the most reliable signal for when to move it. Have them pull it from their pipeline each quarter:

• Closing above 70% of qualified proposals → too cheap. Raise prices.
• Closing below 30% → fix positioning or lead quality, not price.
• Target band: 45–55% close on qualified proposals.

One precondition: this assumes a qualified pipeline of comparable prospects. If most of your work comes through referrals or warm intros, your close rate will run high for reasons that have nothing to do with price — don't raise on that signal alone. And separate lost on price from lost on fit; only the first tells you anything about your number.

When the signal says raise: go up 10–20% at a time and grandfather existing clients. A close rate that's too high isn't a trophy — it's money left on the table. (Sensible defaults; tune them to your own sales motion.)

---

Guardrails (the honesty layer)

• Don't inflate the value number. A capture rate is only as trustworthy as the value it's a percentage of. Conservative, sourceable inputs only.
• Show your work. The method's power is transparency — hand the client the math.
• Respect the market ceiling. Value justifies charging at the top of the band, not above physics. Ignoring the ceiling reads as greed and loses deals.
• Not everything quantifies cleanly. For pure-brand or creative work, use the method to frame the conversation even when the number is a range.
• Never use fake scarcity or inflated "$X value" claims. Specificity beats superlatives; real numbers beat hype.

---

What to output

End with a clear, structured recommendation:

1. The transformation (one line)
2. Year-1 Value Created (with the 3-bucket table)
3. Recommended price / range, with the capture rate and value-to-fee ratio
4. Market check — the band, and where this sits in it
5. Three-tier packaging (project / retainer / productized) with numbers
6. How to present it — the value-first framing line the user can say/write
7. A close-rate note — what to watch next quarter

Then offer the byproduct (don't force it):

"Want me to save your answers as a pricing-profile.md? Next time you price something, I'll pull from it instead of asking again."

If yes, write a short pricing-profile.md with: what they sell, their ICP, their typical transformation, their proof level, their floor, and any known comps. Keep it lean — it's a pricing profile, not a full brand doc, though it can later seed a fuller product-marketing.md.

A note on where it lives: in Claude Code, the user can keep this in ~/.claude/ so it persists across sessions. In Cowork, the sandbox is per-session, so tell them to download the file and re-attach it next time (or paste it back) — it won't automatically persist.

---

Method: "Price on the Proof" by Chris Gee. Builds on the MIT-licensed pricing skill and product-marketing.md pattern by Corey Haines (github.com/coreyhaines31/marketingskills). Willingness-to-pay via the Van Westendorp method. MIT licensed — free to use and adapt.